feat(e2e): Add E2E automated test coverage for PI, Hermas, Openclaw, OpenCode modules

[dongjiang1989]

Background

Currently, the four core modules (PI, Hermas, Openclaw, OpenCode) lack dedicated end-to-end automated test suites running on Kind and agent-sandbox environments. All regression verification depends on manual testing in local or formal environments. This approach brings unstable test environments, heavy operational costs, incomplete scenario coverage, and potential regression risks during daily iteration, code merging and version release.
As module functions and cross-component interactions continue to evolve, manual verification can no longer efficiently and stably verify service availability in containerized scenarios. To solve environment inconsistency, heavy manual workload and unguaranteed regression quality, we need to build a lightweight, repeatable E2E test system based on Kind Kubernetes clusters and agent-sandbox isolation environments, and integrate it into CI/CD workflow for continuous quality inspection.

Problem Description

• PI, Hermas, Openclaw and OpenCode modules have no E2E automated tests validated under Kind + agent-sandbox container environment
• Relying entirely on manual regression testing leads to low iteration efficiency and repeated labor costs
• Lack of isolated sandbox test environment results in inconsistent test results between local and CI
• No automated pre-merge verification, making service regression defects hard to detect at early stage

Solution

Build a full E2E automated test solution based on Kind lightweight Kubernetes cluster and agent-sandbox isolation environment. Add complete test cases for PI, Hermas, Openclaw and OpenCode core workflows and interactive scenarios, support one-click local verification and CI automatic execution, and ensure test consistency, isolation and repeatability.

• Implement E2E test cases covering core business and interactive scenarios for PI, Hermas, Openclaw and OpenCode
• Complete Kind cluster deployment configuration and agent-sandbox environment isolation adaptation for E2E testing
• Unify environment initialization, test execution and resource cleanup specifications for containerized tests
• Support local Kind sandbox debugging and automatic trigger execution in CI pipelines
• Generate standardized test reports to support iterative regression verification and quality tracking

Scope of Changes

• Add E2E automated test cases for the PI module
• Add E2E automated test cases for the Hermas module
• Add E2E automated test cases for the Openclaw module
• Add E2E automated test cases for the OpenCode module
• Add Kind cluster deployment scripts, agent-sandbox isolation config and E2E test runtime scripts

Acceptance Criteria

• All E2E test cases run stably and pass in local Kind + agent-sandbox environment with no flaky failures
• All core processes of PI, Hermas, Openclaw and OpenCode are fully covered without missing key scenarios
• CI pipeline can automatically initialize Kind cluster and agent-sandbox environment, trigger E2E tests and output complete results
• Automated test cases can complete rapid regression verification after code changes, replacing repetitive manual regression work
• The changes will not affect the original business functions, build compilation and other existing test processes